Yotaro Watanabe

Statement map: reducing web information credibility noise through opinion classification

On the Internet, users often encounter noise in the form of spelling errors or unknown words, however, dishonest, unreliable, or biased information also acts as noise that makes it difficult to find credible sources of information. As people come to rely on the Internet for more and more information, reducing this credibility noise grows ever more urgent. The STATEMENT MAP projects goal is to help Internet users evaluate the credibility of information sources by mining the Web for a variety of viewpoints on their topics of interest and presenting them to users together with supporting evidence in a way that makes it clear how they are related. In this paper, we show how a STATEMENT MAP system can be constructed by combining Information Retrieval (IR) and Natural Language Processing (NLP) technologies, focusing on the task of organizing statements retrieved from the Web by viewpoints. We frame this as a semantic relation classification task, and identify 4 semantic relations: [AGREEMENT], [CONFLICT], [CONFINEMENT], and [EVIDENCE]. The former two relations are identified by measuring semantic similarity through sentence alignment, while the latter two are identified through sentence-internal discourse processing. As a prelude to end-to-end user evaluation of STATEMENT MAP, we present a large-scale evaluation of semantic relation classification between user queries and Internet texts in Japanese and conduct detailed error analysis to identify the remaining areas of improvement.

Leveraging Diverse Lexical Resources for Textual Entailment Recognition

Since the problem of textual entailment recognition requires capturing semantic relations between diverse expressions of language, linguistic and world knowledge play an important role. In this article, we explore the effectiveness of different types of currently available resources including synonyms, antonyms, hypernym-hyponym relations, and lexical entailment relations for the task of textual entailment recognition. In order to do so, we develop an entailment relation recognition system which utilizes diverse linguistic analyses and resources to align the linguistic units in a pair of texts and identifies entailment relations based on these alignments. We use the Japanese subset of the NTCIR-9 RITE-1 dataset for evaluation and error analysis, conducting ablation testing and evaluation on hand-crafted alignment gold standard data to evaluate the contribution of individual resources. Error analysis shows that existing knowledge sources are effective for RTE, but that their coverage is limited, especially for domain-specific and other low-frequency expressions. To increase alignment coverage on such expressions, we propose a method of alignment inference that uses syntactic and semantic dependency information to identify likely alignments without relying on external resources. Evaluation adding alignment inference to a system using all available knowledge sources shows improvements in both precision and recall of entailment relation recognition.

A Pipeline Approach for Syntactic and Semantic Dependency Parsing

This paper describes our system for syntactic and semantic dependency parsing to participate the shared task of CoNLL-2008. We use a pipeline approach, in which syntactic dependency parsing, word sense disambiguation, and semantic role labeling are performed separately: Syntactic dependency parsing is performed by a tournament model with a support vector machine; word sense disambiguation is performed by a nearest neighbour method in a compressed feature space by probabilistic latent semantic indexing; and semantic role labeling is performed by a an online passive-aggressive algorithm. The submitted result was 79.10 macro-average F1 for the joint task, 87.18% syntactic dependencies LAS, and 70.84 semantic dependencies F1. After the deadline, we constructed the other configuration, which achieved 80.89 F1 for the joint task, and 74.53 semantic dependencies F1. The result shows that the configuration of pipeline is a crucial issue in the task.

Discriminative learning of first-order weighted abduction from partial discourse explanations

Abduction is inference to the best explanation. Abduction has long been studied in a wide range of contexts and is widely used for modeling artificial intelligence systems, such as diagnostic systems and plan recognition systems. Recent advances in the techniques of automatic world knowledge acquisition and inference technique warrant applying abduction with large knowledge bases to real-life problems. However, less attention has been paid to how to automatically learn score functions, which rank candidate explanations in order of their plausibility. In this paper, we propose a novel approach for learning the score function of first-order logic-based weighted abduction [1] in a supervised manner. Because the manual annotation of abductive explanations (i.e. a set of literals that explains observations) is a time-consuming task in many cases, we propose a framework to learn the score function from partially annotated abductive explanations (i.e. a subset of those literals). More specifically, we assume that we apply abduction to a specific task, where a subset of the best explanation is associated with output labels, and the rest are regarded as hidden variables. We then formulate the learning problem as a task of discriminative structured learning with hidden variables. Our experiments show that our framework successfully reduces the loss in each iteration on a plan recognition dataset.

Organizing Information on the Web through Agreement-Conflict Relation Classification

The vast amount of information on the Web makes it difficult for users to comprehensively survey the various viewpoints on topics of interest. To help users cope with this information overload, we have developed an Information Organization System that applies state-of-the-art technology from Recognizing Textual Entailment to automatically detect Web texts that are relevant to natural language queries and organize them into agreeing and conflicting groups. Users are presented with a bird’s-eye-view visualization of the viewpoints on their queries that makes it easier to gain a deeper understanding of an issue. In this paper, we describe the implementation of our Information Organization System and evaluate our system through empirical analysis of the semantic relation recognition system that classifies texts and through a large-scale usability study. The empirical evaluation and usability study both demonstrate the usefulness of our system. User feedback further shows that by exposing our users to differing viewpoints promotes objective thinking and helps to reduce confirmation bias.

Multilingual Syntactic-Semantic Dependency Parsing with Three-Stage Approximate Max-Margin Linear Models

This paper describes a system for syntactic-semantic dependency parsing for multiple languages. The system consists of three parts: a state-of-the-art higher-order projective dependency parser for syntactic dependency parsing, a predicate classifier, and an argument classifier for semantic dependency parsing. For semantic dependency parsing, we explore use of global features. All components are trained with an approximate max-margin learning algorithm. In the closed challenge of the CoNLL-2009 Shared Task (Hajic et al., 2009), our system achieved the 3rd best performances for English and Czech, and the 4th best performance for Japanese.